AUSTIN AND WITTGENSTEIN ON "DOUBT" AND "KNOWLEDGE" THE imarquette University notion of "doubt" has traditionally played an important role in the philosophy of knowledge. As Descartes stated in his Discourse 011 veth hod, the first rule in seeking truth is never to accept anything unless it is presented clearly and distinctly without any reason or occasion for doubt. Further, even before Descartes, the Platonic conception of knowledge was linked with the very notions of infallibility and unchangeability. Indeed, a general rule for traditional philosophy has been as follows: ij one can doubt the proposition "x is y," then, one cannot say that he has knowledge that "x is y." However, this particular rule has produced puzzlement for philosophers. This puzzlement is called the "problem of knowledge," i.e., Is knowledge possible? If a huinan being is not omniscient or infallible, then there is always a possibility that one can be mistaken or proven wrong by future evidence, and if there always exists such possibility, then there is always grounds for doubting any claim to knowledge. Thus, no claim to possess knowledge can be substantiated. Further, if it is contended that knowledge does not require freedom from such doubt-in other words, if certainty is no longer a requirement of knowledge-then, the very basis for probable statements becomes in jeopardy, for one always claims to know that "x is only probably y." The puzzlement is complete. If there is to be such a thing as knowledge, there must be a human that is not capable of error or correction by future events, but if this is so, then there need be no philosophy of knowledge because infallible, omniscient creatures need have no such concern!! Where does philosophy turn? If certainty is not maintained as a condition of there being knowledge, then no other cognitive claim makes sense. If,
52 F23ASON PAPERS NO. I certainty is maintained as a condition for knowledge, then how can one deny the real occasions for doubt that fallible and limited human beings suffer? It is to this alleged dilemma that Wittgenstein and Austin's conceptions of "knowledge" and "doubt" can be applied. In this paper I will explicate their conceptions and show how their vieus offer the beginning of a solution to this puzzle. Austin's conception of the entire enterprise begins by challenging the initial assumption of the Cartesian and Platonic tradition. Austin writes: Now, we are perfectly aware, and should be candidly, aware of this liability.... The human intellect and senses are, indeed, inherently fallible and delusive, but not by any means inveterately. klachines are inherently liable to break down, but good machines don't (often). It is futile to embark 072 a 'theory of knowledge' which denies this liability: such theories constantly end up by admitting the liability after all, and denying the existence ~J'knowledge.'~ (emphasis added) One must take as the starting point for his epistemology the fact of human fallibility and ignorance. Indeed, there would be no need for an epistemology if this were not the case. One must remember that the goal is to explain what knowledge is for such a being, not for God or anything else. It is to ignore a most elementary of facts to attempt to describe knowledge and show its possibility without Jirst realizing that knowledge is human knowledge. Since the acknowledgement of the fallibility of human consciousness is Austin's starting point in explaining "I know," then the mere fact that it is possible for me to be mistaken is not ground for saying that "I may be mistaken." [Bleing aware that you may be mistaken doesn't mean merely being aware that you are a fallible human being; it means that yozr haue some concrete reason to suppose that you may be mistaken in this case.2 (emphasis added) Thus, since knowledge is human knowledge, it is always possible to be mistaken, but, epistemically speaking, this is an utterly useless type of possibility. When philosophers use such a
"DOUBT" AND "KNOWLEDGE" 53 possibility as a doubt to challenge one's claim to know that I< x is y," there is no disputation of the evidence used to support the knowledge claim but rather, as Austin states, "a challenge as to the reliability" of one's evidence. Yet, every challenge to the reliability of one's evidence rests on the human possibility of error (E. g. Are you sure that you are not dreaming? Is that the right "y"? Etc.), and, as already said, this is epistemically worthless doubt. To clarify this point, the following distinction between two types of possibility will, I think, help us to understand A~stin:~ metaphysical-means given the nature of existence X can occur e.g. It is possible for me to kill you. epistemological-means that there is evidence that X will occur e.g. It is possible that I (a wanton killer) will kill you. Further, I think we can see that it is $so facto invalid to infer epistemological possibility from metaphysical possibility. For example, It is possible for Ghandi to murder. (He has the physical capacity.) Therefore, it is possible that Ghandi will murder. (We have evidence that he is going to do so.) This distinction shows very clearly the point Austin is trying to make-namely, epistemic possibility requires that there be some evidence. We see that it is invalid to doubt the claim to know that "x is y," simply because one can be in error or ignorance. Doubt must be shown, not just asserted. Doubt which is based on the fact that a human can be in error is either not doubt or, rather, nothing other than a requirement for there being knowledge, i.e., the fact that I can be wrong must be there for one to claim that there is knowledge-the "can" shows that it is human knowledge. Thus, to doubt that: I know that "x is y" solely because I can be wrong is to say nothing other than one doubts "x is y" because I am a human being! (This may properly be a conclusion of an investigation into the issue but not an initial assumption.) Besides requiring there to be evidence for there to be doubt, Austin holds that the claim to know that "x is y" is not "predictive" in such a way that the future can always prove it wrong.
54 REASON PAPERS NO. I In other words, the claim that "x is y" will not be proven wrong if circumstances change such that "y" becomes "z"; rather, "'what the future can always do, is to make us rezlise our ideas...."4 Implied here is a position which I would designate as co?ztextualisrn-the view that the truth (rightness) or falsity (wrongness) of a position is a!ways determined in some context of knowledge. To say the same thing in a different way, Austin acknowledges that human knowledge is necessarily limited; thus, to hold that heretofore unknown circumstances prove previous knowledge claims false is wrong-headed. It forgets that knowledge is an activity5 not a static, timeless snapshot. Indeed, the main force of Austin comments on "I know" seem to be that no idea can be produced concerning "x's", goldfinches, or anything else which precludes the ~ossibility that it may be revised. To repeat, newly discovered facts do not, strictly speaking, solely prove false old ideas but, rather, revises them. As said, the claim to know that "x is y" is an activity not a snapshot; it can change to include "and sometimes z" without being made mistaken. The following lengthy quotation characterizes how Austin views the revision of ideas. First, it is arranged that, on experiencing a complex of features C, then we say "This is C" or "This is a C." Then subsequently, the occurrence either of the whole of C or of a significant and characteristic part of it, on one or many occasions, accompanied or followed in definite circumstances by another special and distinctive features, which makes it seem desirable to revise our ideas: so that we draw a distinction between "This looks like a C, but in fact is only a dummy, etc." and "This is a real C (live, genuine, etc.)." Henceforward, we can only ascertain that it's a real C by ascertaining that the special features or complex of features is present in the appropriate circumstances. The old expression "This is a C" will tend heretofore to fail to draw any distinction between "real, live, etc." and "dummy, stuffed, etc." If the special distinctive feature is one which does not manifest itself in any definite circumstances (on application of some specific test, after some lapse of time, etc.) then it is not a suitable feature on which to base a distinction between "real" and "dummy, imaginary, etc." All we can then do is to say "Some C's are and some aren't, some do afid some don't: and it may be very interesting or important whether they do or don't, but they're all C's, real C's just the same. Now if the I I
special feature is one which must appear in (more or less) definite, circumstances, then "This is a real C" is not necessarily predictive: we can, in favourable cases, make sure of it.6 (The distinction could just as easily be between C's and D's as C's and real C's.) This illustrates most aptly that newly discovered facts, especially facts that do not fit into previous conceptual categories, are not a threat to knowledge but an expansion of it. Austin's comments concerning "doubt" and "knowledge" fit together quite nicely. "Doubt" can never be solely based on the possibility that one can be wrong, and "I knowj' is never so static as to be proven false by merely the new discovery of facts. Both comments are based on Austin's initial declaration that the starting point of an epistemology is the recognition that knowledge is human knowledge. Consider the proposition "I know that x is a tree." It [the proposition] would not be surmise and I might te!i it to someone else with complete certainty, as something there is no doubt about. Bur does that mean that it is unconditionally the truth? May not the thing that I recognize with complete certainty as the tree that I have seen here my whole life long-may this not be disclosed as something different? May it not confound me? And nevertheless it was right, in the circumstances that give the sentence meaning, to say 'I know (I do not merely surmise) that that's a tree.' To say that in strict truth I only believe it, would be wrong.... I cannot be making a mistake about it. But this does not mean that I am infallible about it.' (emphasis added) Here Wittgenstein echoes Austin's claim (Or is it vice-versa?) that "I know" cannot be "predictive" in such a way that the future can prove it wrong. He is clearly contending that "I know" does not in any way amount to a claim of infallibity. The claim "I know that x is a tree" is justified within the circumstances that give the sentence meaning. (It should be noted that the' "circumstances that give the sentence meaning" is the language-game or context in which the sentence is found. For Wittgenstein this is "rock-bottom''
56 REASON PAPERS NO. I or the "point where explanation ends." We shall have need to keep this in mind.) The contextuality implied here is explicitly brought out in the following remarks: That to my mind someone else has been wrong is no ground for assuming that I am wrong now.-but isn't it a ground for assuming that I might be wrong? It is no ground for any unsureness in my judgement, or my action^.^ I act with complete certainty. But this certainty is my This shows that "I know" is always used in a context and that it is always someone's "I know." To ignore this is to ignore the language-game in which it is found. For example, considering the proposition, "I know that I have never been on the moon," Wittgenstein states,... even the thought that I might have been transported there, by unknown means, in my sleep, woz~ld not give me any right to speak of a possible mistake here. Iplay the game wrong if I do.lo (emphasis added) What is the wrong move made? It is confusing an imagined doubt with a real doubt, for one should not... say that one is in doubt because it is possible for one to imagine a doubt. I can easily imagine someone always doubting before he opened his front door whether an abyss did not yawn behind it... but that does not make me doubt in the same case.ll Indeed, "one gives oneself a false picture of doubt";12 one needs a grounds for doubt-a reason found in the circumstances surrounding the claim "I know x is y."13 Doubt is only found in various contexts (language-games) and, thus, its ground is dictated accordingly. Yet, if it is maintained that there is a sense of "doubt" that can be applied to the proposition "I know x is y" because one is truly a fallible human being. I think Wittgenstein meets the objection by granting it but showing it to be epistemically useless. His question: "Can one say: 'Where there is no doubt there is no knowledge either'?"i4 points out this approach. Sure, one car2 be mistaken, for to be human, at least, admits to that possibility, but what of it? How does this show that one might be in error in the situation where one claims' to know that
x is y? Such doubt is not dictated by the language-game. Thus, for Wittgenstein such a possibility of error does not effect the epistemic worth of the claim to know that x is y. Wittgenstein admits that it is quite possible for one to say that "I know x is y" within a context or language-game and, then, have the language-game alter in such a way that doubt is introduced regarding the claim. This, however, would only mean that the language-game changed.15 The original claim that "I know x is y" in its context is (was) correct nonetheless. (Remember a proposition is meaningsless outside of its languagegame.) Thus, I think, this is what enables Wittgenstein to say, "I have the right to say, 'I can't be making a mistake about this' even if I am in error."16 One's claim to know that x is y is, thus, not proven wrong by the new language-game (context) surrounding it; rather, the meaning of the claim changes. A claim of certainty may be reduced to a claim of probability, for example. As said, the correctness of each claim is maintained in its context, despite the fact that a!teraticn occurs, for "the concept of knowing is coupled with that of language-game."17 This is further pointed out when we say we know that water boils and does not freeze under such-andsuch circumstances. Is it conceivable that we are wrong? Wouldn't a mistake topple all judgement with it? More, what could stand if that were to fall? Might someone discover something that made us say 'It was a mistake'? Whatever may happen in the future, however water my behave in the future-we know that up to now it has behaved thus in innumerable instances. This fact is fused into the foundation of our language-game.ls Thus, if tommorrow a new discovery proves water not to boil at roo0 C. at sea level, this does not in the least effect the previous knowledge claim-one can claim that old knowledge was expanded or revised but not mistaken, for the new discovery only alters the language-game. (Wittgenstein, however, would not like to say that "facts" alter language-games, but he would say the language-games alter and that's the important point here.) Further, to say that human knowledge is contextual (within a language-game) is to say something unnecessary. As Wittgen-
58 REASON PAPERS NO. I stein states, "A judge might even say 'That is the truth-so far as a human being can know it.' But what does this rider accomplish."?i9 (emphasis added) ltery simply then, Wittgenstein seems to hold that knowledge is always determined within a context, a language-game, and that real doubt must be grounded there also and not just imagined or thought. Austin's and Wittgenstein's conceptions of "knowledge9' and "doubt" amount to the following points: I) Goivledge is not a timeless snapshot rather it is an activity. 2) A11 knowledge is contextual. 3) Doubt h ~ to s he grolunded in evidence beyond the knower's humanity. 4) Context is fundamental in determining the epistemic worth of a proposition. 5) All knowledge is human knowledge. Strange as it may seem, the fifth point is the most important realization, for it immediately points out that knowledge is not intrinsic to the world; rather, knowledge depends on human activity to exist-meaning is not found in things but with the activity of humans with things. (This is not meant to imply in anyway that knowledge is "subjective" in the sense of arbitrary; rather, it simply tries to show human activity as a necessary condition for knowledge.) This realization makes all the other points possible, for all of the other points start with the fact of human knowledge. Since human knowledge occurs for creatures that are not omniscient, knowledge must be subject to alteration and not something timelessly static. However, since human knowledge still requires certainty (as seen, "probability" requires certainty), knowledge claims must be found in a context and, further, the context must be the basis from which truth, falsity, correctness, incorrectness, and rightness and wrongness are
determined-no proposition is a a-contextual, even this one! Finally, since human knowledge cannot exist unless it can be mistaken, doubt must be based on evidence that something is not the case. As said, all these points follow directly from the fact that knowledge is human knowledge. How does all this solve the "problem of knowledge"? It solves the problem by showing that the metaphysical possibility of error or correction by future events does not constitute evidence for doubting the claim that "one knows that x is y"; rather, there must be a concrete reason to doubt the claim. This, of course, only eliminates the constant doubt. It is the contextual and active nature of knowledge that makes certainty possibleone may still be tempted to say contextual or human certainty, but this temptadon ca11 be squashed by asking: As opposed to what? (The fundamentality of the language-game jumps right up!!) It may be objected that this view of "knowledge" and "doubt" proves too much, for does it ever ailow for someoiie to be ix error? 1f knowledge is always expanded and revised by new discoveries, are we not just saying that one is never wrong? Thus, haven't we just substituted one extreme with another? This objection is a good one and much is required by way of answering, more than can be supplied in this paper. However, I think Wittgenstein has the key element in the answer. He says, "There is a difference between a mistake for which, as it were, a place is prepared in the game, and a complete irregularity that happens as an exceptj~n."~~ I take this to mean that errors, mistakes and other assorted blunders occur within a context such that one can point out that the rules are not being followed and, thus, point out mistakes. Complete irregularities are outside of the context, and one doesn't know what to say about them. Thus, this view would still allow for errors but would not let "complete irregularities" destroy the possibility of knowledge. However, it is not at all clear to me how one can say which is a "mistake" within a context and which is a "complete irregularity," for that seems to depend solely on who is noticing the occurrence. In other words, it would seem that the more knowledgeable person concerning the language-game would
60 REASON PAPERS NO. I consider more occurrences mistakes than the less informed person. However, possibly this is not a damaging result, for Wittgenstein does say that the complete certainty is my certainty. This does not mean that the determination of an occurrence as a "mistake" or "complete irregularity" is entirely arbitrary; rather, this would have to be done by reference to all the known data concerning the occurrence. If in the widest context of knowledge, this occurrence could have been prevented, then a "mistake" has occurred; if in the same context, there was no data that could forecast such an occurrence, then, a "complete irregularity," such as Austin's exploding goldfinch, has occurred, and we just can't say anything about that. In this way, then, error can be allowed for without doing away with kno~vledge.21 1 J. L. Austin, "Other Minds," Philosophical Papers, 2nd ed. (London: Oxford University Press, 1970)~ p. 98. Ibid. Z Waiien I from Jarret 9. Wollstein's "N~tes from 'Certain?? 7,Vitho-t Omin- science'," notes of a lecture given by Dr. Leonard Peikoff at the University of Virginia in 1967. *Austin, pp. 88-89. Ibid., p. 103. Ibid., p. 89. ' Ludwig Wittgenstein, On Certainty, trans. Anscombe and Wright (Oxford: Basil Blackwell, 1969), p. 55 e. Ibid., p. So e. Ibid., p. 25 e. lo Ibid., p. 88 e. Wittgenstein, Philosoplzical Investigations, trans. Anscombe (New York: McMillan Co., 1953), p. 39 e. l2 Wittgenstein, On Certainty, p. 33 e. ls Ibid., pp. 33 e-34 e. l4 Ibid., p. 18 e. l6 Ibid., p. 85 e. Ibid., p. 88 e. l7 Ibid., p. 74 e. Is Ibid., p. 73 e. Is Ibid., p. 80 e. 2D Ibid., p. 85 e. " This paper has only attempted to give some important characteristics con<erning knowledge claims and not a definitive answer as to what knowledge is.